PROJECT TITLE :

Adaptive Graph Signal Processing: Algorithms and Optimal Sampling Strategies - 2018

ABSTRACT:

The goal of this Project is to propose novel strategies for adaptive learning of signals outlined over graphs, which are observed over a (randomly) time-varying subset of vertices. We tend to recast two classical adaptive algorithms within the graph Signal Processing framework, specifically the smallest amount mean squares (LMS) and the recursive least squares (RLS) adaptive estimation ways. For each methods, an in depth mean-sq. analysis illustrates the effect of random sampling on the adaptive reconstruction capability and therefore the steady-state performance. Then, many probabilistic sampling ways are proposed to style the sampling likelihood at every node within the graph, with the aim of optimizing the tradeoff between steady-state performance, graph sampling rate, and convergence rate of the adaptive algorithms. Finally, a distributed RLS strategy is derived and shown to be convergent to its centralized counterpart. Numerical simulations administrated over both artificial and real information illustrate the nice performance of the proposed sampling and recovery methods for (distributed) adaptive learning of signals outlined over graphs.